python多线程threading使用Semaphore或BoundedSemaphore实现并发限制

信号量：

class threading.Semaphore([value]) values是一个内部计数，values默认是1，如果小于0，则会抛出 ValueError 异常，可以用于控制线程数并发数

1.acquire([blocking]) 2.release()

例如： 可用用来限制数据库的连接数

maxconnections = 5
...
pool_sema = BoundedSemaphore(value=maxconnections)

pool_sema.acquire()
conn = connectdb()
... use connection ...
conn.close()
pool_sema.release()

例如：

# encoding: UTF-8
import threading
import time

def showfun(n):
    print "%s start -- %d"%(time.ctime(),n)
    print "working"
    time.sleep(2)
    print "%s end -- %d" % (time.ctime(), n)
    semlock.release()

if __name__ == '__main__':
    maxconnections = 5
    semlock = threading.BoundedSemaphore(maxconnections)
    list=[]
    for i in range(8):
        semlock.acquire()
        t=threading.Thread(target=showfun, args=(i,))
        list.append(t)
        t.start()

结果：

Sat May 27 02:43:00 2017 start -- 0
working
Sat May 27 02:43:00 2017 start -- 1
working
Sat May 27 02:43:00 2017 start -- 2
working
Sat May 27 02:43:00 2017 start -- 3
working
Sat May 27 02:43:00 2017 start -- 4
working
Sat May 27 02:43:02 2017 end -- 0
Sat May 27 02:43:02 2017 end -- 1
Sat May 27 02:43:02 2017 end -- 2
Sat May 27 02:43:02 2017 end -- 3
Sat May 27 02:43:02 2017 end -- 4
Sat May 27 02:43:02 2017 start -- 5
working
Sat May 27 02:43:02 2017 start -- 6
working
Sat May 27 02:43:02 2017 start -- 7
working
Sat May 27 02:43:04 2017 end -- 5
Sat May 27 02:43:04 2017 end -- 6
Sat May 27 02:43:04 2017 end -- 7

线程数限制到5个，因此等待0-4完毕之后5-7才能请求到资源进行执行